Recently, there has been an increased interest in "individualized medicine" and thus pharmacogenetics and pharmacogenomics in cancer research have moved to the forefront as well. Pharmacogenetics is the study of the role of inheritance in individual variation in response to drugs, nutrients and other xenobiotics. In this post-genomic era, pharmacogenetics has evolved into pharmacogenomics, a discipline that has been heralded as one of the first major clinical applications of the striking advances that have occurred and continue to occur in human genomic science. Statistical methods that combine various sources of genomic data would likely provide novel insights, yet are lacking in pharmacogenomic studies. Joint analysis of multiple types of genomic data is advantageous, especially when the etiology of the disease or phenotype is complex. For example, my group has been collaborating with a pharmacogenomic research group that is conducting studies of the effect of anti-cancer drugs on lymphoblastoid cell lines (Coriell Human Variation Panel) to begin to define the effect of common genetic variation on drug response phenotypes. These studies have included cytotoxicity measures at multiple drug concentrations, basal mRNA expression array data, post-drug treatment mRNA expression data, metabolite data, resequencing data for genes in known drug pathways, and genome-wide genetic information in the form of single nucleotide polymorphisms (SNPs), and copy number variation (CNVs). Analysis of this data-rich "model system" is a challenge. A statistical approach that could account for disparate "layers" of genomic data is based on Bayesian methodology. Over the last few decades, applications of Bayesian methods by Markov Chain Monte Carlo (MCMC) have increased with the advancement of computers and computational methods, particularly with application to genetic data. In this current grant application, we intend to combine various sources of phenotypic and genomic data collected on the "Human Variation Panel" cell lines (e.g. cytotoxicity data, mRNA expression data, genotypic data) into pharmacogenomic models using Bayesian methods that will assist researchers in generating hypotheses that will lead to better understanding of the complex nature of the relationship between the genome and drug response, leading eventually to the development of "individualized therapy" for cancer patients. To accomplish this, we will be combining path modeling ideas into a Bayesian hierarchical nonlinear model to assess the relationship between cytotoxicity drug endpoints and genomic information. PUBLIC HEALTH RELEVANCE: This application proposes to develop novel statistical methods for the joint analysis of genomic data collected on the Coriell Human Variation Panel cell lines involving pharmacogenomic studies of anti-cancer drugs. In addition, results from these pharmacogenomic studies involving the cell lines will be tested in pharmacogenomic translational studies, using genomic information collected from cancer patients treated with the anti-cancer drug. These translational studies will test whether genetic variations identified from the pharmacogenomic studies of the cell lines might be associated with clinical response. Therefore, these statistical methods will have broader applications in pharmacogenomic studies beyond cell-based model systems to translational studies involving cancer patients treated with chemotherapeutic agents. These models will also aid investigator in generate hypotheses that will lead to better understanding of the complex nature of the relationship between genomic variation and drug response, leading eventually to the development of "individualized therapy" for cancer patients. These models will also be applicable to the study of complex diseases where multiple types of genomic data are collected.